Mean values when the dot model is not the most supported

questions concerning analysis/theory using programs M-SURGE, E-SURGE and U-CARE

Mean values when the dot model is not the most supported

Postby simone77 » Fri Nov 25, 2011 7:09 am

Hi,

I want to see if survival rate varies according to their departure state. AIC supports this (AIC {phi= f.t (GEMACO)} << AIC {phi=t}), but the time varying model (phi= f.t) estimates of phi have huge CIs (due to data sparseness).
Still, I would like to get mean values of phi depending on their departure state to get an idea on how much they differ among them.
The dot model has quite acceptable CIs but I read that is not the best way to calculate this in the MARK Gentle Introduction (pg 6-83) where they suggest an alternative approach based on modifying the Design Matrix.
I have tried to do the same in the Constraint Matrix in the Gemaco Interface but I am a bit lost because the code is not as I would expect according to my past experience with MARK interface and because I don't know how to modify that.

Thanks for any help,

Simone
simone77
 
Posts: 197
Joined: Mon Aug 10, 2009 2:52 pm

Re: Mean values when the dot model is not the most supported

Postby CHOQUET » Mon Nov 28, 2011 8:43 am

Hello,

There is one easy way to build the mean and variance of the mean by the delta method.

mean(phi)=sum phi(t)/T

var(mean(phi))=1/(T*T) transpose(1_T) VAR(phi) 1_T

where 1_T is a vector of one of size T and VAR(phi) is the block matrix (of size T) of the var-cov matrix
given by E-SURGE.

Rémi
CHOQUET
 
Posts: 211
Joined: Thu Nov 24, 2005 4:58 am
Location: CEFE, Montpellier, FRANCE.


Return to analysis help

Who is online

Users browsing this forum: No registered users and 20 guests